Springer is definitely bad; they're mostly just a service for printing books, stamping their logo on them, and getting them into academic libraries. However, MIT Press is excellent. There's a big range of support and quality control among publishers. Zero Books is also quite successful as a nontraditional quasi-academic press. They've developed a niche style that's popular in a certain audience (academic but readable and short, ~70-100 page books on a specific subject). So authors publishing a "Zero Book" through Zero get a bunch of built-in exposure to the audience the publisher has cultivated, in part through marketing (the brand is known, individual academics often aren't), and in part because readers trust it to do some quality/style control so the books in their catalog fit the expected style.
As a reader I like finding publishers like that because it reduces the noise for me considerably and improves discoverability: browsing http://mitpress.mit.edu/books/new-titles or http://www.zero-books.net/forthcoming-titles.html is actually manageable. With a few such publishers (I follow about 5 semi-regularly) you can keep abreast of a lot of interesting things happening without having to recognize the individual authors or do extensive research on them; I outsource the job of doing that to the publisher.
As a reader I like finding publishers like that because it reduces the noise for me considerably and improves discoverability: browsing http://mitpress.mit.edu/books/new-titles or http://www.zero-books.net/forthcoming-titles.html is actually manageable. With a few such publishers (I follow about 5 semi-regularly) you can keep abreast of a lot of interesting things happening without having to recognize the individual authors or do extensive research on them; I outsource the job of doing that to the publisher.